Net Promoter Score has served the industry very well as the standard in satisfaction intention, and can now be enhanced by adding explicit customer satisfaction data, influence and customer reviews already on the web.
NPS, a Industry Standard Before the Social Web
There’s no better way to measure customer satisfaction and intention to refer than the Net Promoter Score. In fact, this simple mechanism asks consumers on a 0 to 10 rating scale:
“How likely is it that you would recommend our company to a friend or colleague?” Based on their responses, customers can be categorized into one of three groups: Promoters (9-10 rating), Passives (7-8 rating), and Detractors (0-6 rating)
NPS ,while effective at capturing the intention of advocacy, does not measure actual advocacy or detractions that occur in the social web. As a result it’s difficult to capture the entire ‘net’ experience as the social web has demonstrated.
Now The Social Web Provides New Data Sources
The social web actually records customers making explicit ratings, rankings, recommendations or warnings about products in services, I’ve given some pragmatic reasons on why this is important. You can find these reviews in Amazon, Twitter, Plancast, Yelp, Facebook, Twitter, and beyond. In particular, the social web allows brands to actually measure three new types of public data sources:
- Customer Satisfaction: Customers can now provide ratings, reviews, and other critiques in online review sites.
- Influence: Not all customers are created equal, in fact some customers have have great breadth of reach (like Celebrities on Twitter) or have depth in knowledge (expert blogger in your market).
- Referral Activity: No need to ask “how likely” they are to refer, you can see them do it live.
Get Accurate, Measure the Total Social Customer Value (TSCV)
Companies must value both the total customer satisfaction as well as influence and advocacy behaviors in order to provide a holistic example of the modern customer.
Matrix: Know Your Total Social Customer Value with 6 Factorials
| Attribute | Why it’s important | Data Location | How you should measure | What no one tells you |
| Net Promoter Score | This is the mainstay of customer satisfaction measurement and shouldn’t go away. It’s easily understood, well documented, and is a useful metric to overall ‘referral intention’. (Intention doesn’t measure actual behavior, just the likelehood you would) | Support exit surveys, primary research surveys, work with Satmetrics, the owners of this methodology | 1-10 Referral score: Promoters (9-10 rating), Passives (7-8 rating), and Detractors (0-6 rating) | This is the standard default measurement, yet needs additional factorials to represent the modern customer. |
| Influence (Absolute) | To determine if a customer is influential to others, such as celebrities, top bloggers, analysts and media. This doesn’t necessarily mean however they are trusted by your specific market. | There are a variety of secondary sources such as brand monitoring firms, like Buzzlogic, Radian6, as well as reputation management systems like Rapleaf. Your PR firm will have this list of absolute influencers, and their Twitter/blog/RSS numbers are good indicators | Total possible reach, frequency of publication. | These large influencers can cause mainstream media to shift attention, and will impact SEO, but don’t expect your actual consumers to trust them as much as they trust their peers. Assume high scoring at this level is towards the wider part of the funnel and may influence the lower elements. |
| Influence (Relative) | These are individuals that are ‘experts’ in your particular market. While they may not have mainstream appeal, they may influence consumers directly. For example, bloggers that write a dedicated blog to your market, or super reviewers that provide detailed reviews about your products in online sites | Online communities, Technorati data, and brand monitoring firms | Unlike Absolute Influence, we’re looking for depth –not breadth of ability. Look for how detailed, knowledgeable and how much they engage with prospects and consumers. | This measurement must be factored into the overall formula as they have direct influence during the decision making stage in lower funnel. |
| Advocacy (Intention) | Data that indicates a prospect is ‘willing’ to purchase, but has not yet. The difference here is that they do so in public. | Wish lists, shopping carts, or intention based data sources like Plancast, Facebook Events, Tripit, Dopplr. | This data is difficult to get, as it’s currently not aggregated. Expects Social CRM systems to emerge that will help to assemble all this data around a single profile. | Intention data is has high potential and therefore value, but low accuracy. Expect to factor in the chance that individuals will not move to the next stage, so conduct weighted averages here on probability. |
| Advocacy (Purchase, or Post Purchase) | This is the most key measure, as it measures when customers actually explicitly share with others that they have purchased a product, and may have posted an opinion, influencing others. | A variety of locations like Twitter, review sites, blogs, and social networks. See how vertical based review sites are emerging like GDGT, where consumers share their technology products with their peers, influencing purchase behavior. | Since you’ve already factored in their influence from above, you’ll add sentiment and accuracy. | The key that most marketers don’t realize is that here’s the opportunity to be proactive. If there are positive reviews, figure out how to broadcast them to other channels, see how Bazaarvoice and Zuberance help brands with this. For negative reviews, the company can contact the negative reviewed to provide updates to service or product and request a second review. |
| Referral Activity | The absolute measure if a single individual or community has caused others to buy. | Referral codes in eCommerce systems, or surveys at point of sale, or special tracking tools from existing web analytics tools (cookies, 1X1 pixels, etc) | Provide advocates with referral codes, so they can encourage their friends to buy, or special tracking features to ensure an accurate measurement. | The key is to look for the path of least resistance among the social channels to identify where referral activity flows smoothes |
| SUM | Total Social Customer Value attempts to measure the entire value of customer satisfaction, influence, and advocacy in both intention and historical data types | While still early, expect this data to be collected into a social CRM system, then be exported to business intelligence software systems like SAS, Qlikview, Oracle, SAP, Microstrategy, and others | Adding all these factorials will develop a more accurate view of social customers | While early and mainly theory, I know of a few brands that are attempting to scrap this data together and glue it together. Expect a new analytics agency to emerge that will help aggregates this. |
How To Use This Formula
The above matrix is a formula to develop your Total Social Customer Value (TSCV) yet will require modification and customization for each company, as different verticals have higher rates of customer social activity. For example, technology, consumer product goods, and hospitality industry have a high degree of customer interactions –while component parts or specialized industries like tractors may not. You’ll need to develop a formula, use your social CRM system to track this and eventually use business intelligence software that can calculate this.
Total Social Customer Value = Net Promoter Score times Absolute Influence times Relative influence times Advocacy Intention times Advocacy at Point of purchase or post purchase times actual referrals
or
TSCV= NPS x Ia x Ir x Ai x Ap x R
Challenges to this Methodology
My goal was to define how measurement of customer satisfaction, influence and advocacy have changed due to the social web, yet there are few challenges to this methdology, which I’m happy to point out as an industry analyst: 1) This is new thinking and not everyone will get it. 2) Data is disparate and scattered among the disparate web 3) It’s more complex, and requires more time to put together, however the accuracy of the data will be higher. 4) Social CRM is still very early, and aggregating all data into a single area is a challenge.
Relevant Research: Social Analytics and Social CRM
We’ve conducted research and published the following reports on Social Marketing Analytics (slides and recording, and the report) with 3-ex Forrester/Jupiter analysts Social CRM Report, with my colleague Ray Wang which will be the system to house all of this data, along with intelligent filtering and workflows. You’ll need to know more about the formulas to amend NPS, as well as understand the new processes and technologies in Soical CRM which will capture a single user profile regardless of where they publish online.
Praise to the Net Promoter Score
To be clear, I’m not suggesting that we do away with NPS, it still serves a vital function to customer satisfaction, and respect the value it’s brought to the industry. There’s no other single question that can yield so much intention data, It’s a fantastic tool and has advanced the industry, and I know we can build on top of it to reflect the moden web. I’m thankful for the work that Fred Reichheld, Bain & Company, and Satmetrix have done to create Net Promoter Score, and I fall into the promoter score as ‘9’ on the scale, as I’m here recommending you use this methodology, with my additional caveats.
Agreed, it is more complex, but so is the modern day customer. Modern day customers are now influencing each other using the social web, and our measurement systems must eventually adapt to this.
While complex now, different groups within the organization are *already* capturing this information. Corp comm and PR are already tracking influence, and eCommerce is looking at ratings and reviews.
Agreed, it is more complex, but so is the modern day customer. Modern day customers are now influencing each other using the social web, and our measurement systems must eventually adapt to this.
While complex now, different groups within the organization are *already* capturing this information. Corp comm and PR are already tracking influence, and eCommerce is looking at ratings and reviews.
Jeremiah,
Interesting post, but I think you are over-complicating it. Paul Greenberg touched on this as well with some scientifically studied enhancements (http://bit.ly/cSElb7) – I used his post for some discussion here: http://bit.ly/afTcQ5 and Graham Hill also did an analysis (http://bit.ly/cGVlgj) back in November of 2008 (Graham's post is quite thorough as are his comments on my post).
Mitch
access to social measures is not the point… the point is that an arbitrary mash-up of data points combined by some arbitrary formula really has no validity. Who knows how those measure truely relate to each other. I'm sure the relationship is more sophisticated than multiplying them with each other. The reason NPS was created was to provide an unburdened metric of consumer advocacy…..using social can be an elegant no-brainer extension of NPS but not in this bloated way IMHO.
access to social measures is not the point… the point is that an arbitrary mash-up of data points combined by some arbitrary formula really has no validity. Who knows how those measure truely relate to each other. I'm sure the relationship is more sophisticated than multiplying them with each other. The reason NPS was created was to provide an unburdened metric of consumer advocacy…..using social can be an elegant no-brainer extension of NPS but not in this bloated way IMHO.
I met with the ecommerce team at a large online retailer recently who is working on all these formulas now. For them, the more intelligence can increase margins just a few percents, which can result in millions of dollars of change in revenue or cost to reach customers.
For them, they already have teams in place working on this, it's key for them. More on that soon.
I met with the ecommerce team at a large online retailer recently who is working on all these formulas now. For them, the more intelligence can increase margins just a few percents, which can result in millions of dollars of change in revenue or cost to reach customers.
For them, they already have teams in place working on this, it's key for them. More on that soon.
Jeremiah, I am curious why you chose a multiplicative rather than additive formula. Multiplying all factors together means increasing one also increase the impact of all other factors. Any thoughts?
Jeremiah, I am curious why you chose a multiplicative rather than additive formula. Multiplying all factors together means increasing one also increase the impact of all other factors. Any thoughts?
I tried to add a comment here a while back, but it included links, therefore I believe it is being put into moderation. Thus, I will quickly suggest that a quick review of Graham Hill on CustomerThink from November 2008 and Paul Greenberg from August 2009 where Paul referenced “Georgia State University professor Dr. V. Kumar introduced new measurements beyond CLV: customer brand value (CBV) and customer referral value (CRV).” Before we try something new to change NPS (which is flawed, IMHO).
To suggest quantitative measures we need quantitative data to lead us down the right path. For example in Dr. Kumar's work “Intent to refer and ultimate profitability are not strongly correlated.”
I tried to add a comment here a while back, but it included links, therefore I believe it is being put into moderation. Thus, I will quickly suggest that a quick review of Graham Hill on CustomerThink from November 2008 and Paul Greenberg from August 2009 where Paul referenced “Georgia State University professor Dr. V. Kumar introduced new measurements beyond CLV: customer brand value (CBV) and customer referral value (CRV).” Before we try something new to change NPS (which is flawed, IMHO).
To suggest quantitative measures we need quantitative data to lead us down the right path. For example in Dr. Kumar's work “Intent to refer and ultimate profitability are not strongly correlated.”
Thanks Mitch, I didn't know about their work, I'll go approve that other comment.
Thanks Mitch, I didn't know about their work, I'll go approve that other comment.
that makes it right?
that makes it right?
Very good thinking, Jeremiah. I think that it makes a lot of sense and that there's room for an improvement: the final calculation needs to be weighed very well. All the single elements must have their importance relative to the others. As it is now, if I'm not mistaken, having one more referral has the same weight of being a strong influencer for, let's say, another influencer.
Right now it's a great framework to define how brand / product is perceived by the target, it only needs to be a little more precise in the “sum”, IMO.
I'm becoming a big fan of your “matrix” category. 😉
Very good thinking, Jeremiah. I think that it makes a lot of sense and that there's room for an improvement: the final calculation needs to be weighed very well. All the single elements must have their importance relative to the others. As it is now, if I'm not mistaken, having one more referral has the same weight of being a strong influencer for, let's say, another influencer.
Right now it's a great framework to define how brand / product is perceived by the target, it only needs to be a little more precise in the “sum”, IMO.
I'm becoming a big fan of your “matrix” category. 😉
Jeremiah, thanks for contributing to the collective thinking around NPS. It's not an unreasonable argument to make that NPS is all about simplicity, and your approach adds complexity. However, the simplicity factors for NPS mainly have merit in getting organizational alignment and the engagement of top management and front line employees, not the analytics team. Most successful programs have pretty sophisticated analytics and there is no reason that this approach, assuming it creates additional insights, could not add value.
Jeremiah, thanks for contributing to the collective thinking around NPS. It's not an unreasonable argument to make that NPS is all about simplicity, and your approach adds complexity. However, the simplicity factors for NPS mainly have merit in getting organizational alignment and the engagement of top management and front line employees, not the analytics team. Most successful programs have pretty sophisticated analytics and there is no reason that this approach, assuming it creates additional insights, could not add value.
Jeremiah,
This is very timely. Behind the scenes I know several major clients who are grappling with these same issues. NPS is the standard, but that they also know they will need to modify their approach to take into consideration the new social consumer. It seems like we are headed down a path to some sort of extended NPS metric. Brands are willing to create new methodologies to get there. I can see a point down the road when even JD Powers Customer Satisfaction Index will include this new line of thinking.
Mike Spataro
Visible Tech
Jeremiah,
This is very timely. Behind the scenes I know several major clients who are grappling with these same issues. NPS is the standard, but that they also know they will need to modify their approach to take into consideration the new social consumer. It seems like we are headed down a path to some sort of extended NPS metric. Brands are willing to create new methodologies to get there. I can see a point down the road when even JD Powers Customer Satisfaction Index will include this new line of thinking.
Mike Spataro
Visible Tech
Great post. I think the advocacy metrics especially as they relate to the others is hardest to quantify and model out. Influence too unfortunately is a bit of a moving target once you get into the long tail. Still getting the right analytics will only get better in time. You'll probably find the SIM Score I introduced last summer interesting. It borrows from the philosophies of the Net Promoter Score and serves as a brand health metric – originally for the social web but increasingly as a broader indicator. So unlike NPS or TSCV which are about customer satisfaction, this one is brand health. Info on it is here http://www.goingsocialnow.com/2010/06/the-new-b… and here http://fluent.razorfish.com Watch for a new version of it soon incorporating influence etc better.
Let's hold judgement and see over the next few months as brands roll this out. I know it's happening, let's see how it plays out.
A new methodology? It depends.
The more important thing to remember is getting to know customers better, is always a good thing. Thanks G2.
Let's hold judgement and see over the next few months as brands roll this out. I know it's happening, let's see how it plays out.
A new methodology? It depends.
The more important thing to remember is getting to know customers better, is always a good thing. Thanks G2.
Jeremiah,
I intend to write a post very soon that touches on a similar topic but in a slightly different way. A more complete view of customer value is what we are all after.
The Ultimate Question (the driver for NPS) is what it is – it's a high leverage question. It probably gives you the most feedback for the most amount of time invested. It's a great question, but a potentially dangerous one if used incorrectly. (I know that's what you are trying to fill in here).
Thanks again. I look forward to continuing the conversation.
Jeremiah,
I intend to write a post very soon that touches on a similar topic but in a slightly different way. A more complete view of customer value is what we are all after.
The Ultimate Question (the driver for NPS) is what it is – it's a high leverage question. It probably gives you the most feedback for the most amount of time invested. It's a great question, but a potentially dangerous one if used incorrectly. (I know that's what you are trying to fill in here).
Thanks again. I look forward to continuing the conversation.
Jeremiah,
Long time fan, but I feel really let down by this post. The NPS was created as nothing more than a loyalty rating formulated so it was easy to understand by everyone within an organization.
Also, the NPS is nothing but a qualitative measuring stick of correlation that answers the question “How do our customers feel about us?” – which is generally the question asked by Leadership after they review quantitative social media data (ex. revenue, conversion, etc.). The metric we've been relying on over the past couple of years to answer this question is “sentiment” which is filled with its flaws (maybe 70% accurate at best). At least the NPS offers a consistent and reliable measurement that transcends other efforts of a company.
And there lies the rub. This formula that you've proposed is incredibly complicated and just as flawed as the NPS – ex. there is no scientific evidence that a suggested “recommendation” leads to a sale. This same reasoning can be applied to your other variables.
If asked to propose this formula to a client, I would be very uncomfortable trying to explain this concept, much less the (possible) ROI of this effort.
Having worked in Operations Leadership for over 13 years, if a consultant tried to explain this to me as a metric for adoption, I would have stopped them by the second variable as front line employees – the true brand ambassadors of any organization – would not understand it.
I recommend that you stick with just the NPS – it's not perfect, but it's easy to understand by everyone and easily compliments hardcore metrics. Don't try to reinvent the wheel with these efforts.
Cheers,
Mark
Jeremiah,
Long time fan, but I feel really let down by this post. The NPS was created as nothing more than a loyalty rating formulated so it was easy to understand by everyone within an organization.
Also, the NPS is nothing but a qualitative measuring stick of correlation that answers the question “How do our customers feel about us?” – which is generally the question asked by Leadership after they review quantitative social media data (ex. revenue, conversion, etc.). The metric we've been relying on over the past couple of years to answer this question is “sentiment” which is filled with its flaws (maybe 70% accurate at best). At least the NPS offers a easy, consistent and reliable measurement that transcends other efforts of a company.
And there lies the rub. This formula that you've proposed is incredibly complicated and just as flawed as the NPS – ex. there is no scientific evidence that a suggested “recommendation” leads to a sale. This same reasoning can be applied to your other variables.
If asked to propose this formula to a client, I would be very uncomfortable trying to explain this concept, much less the (possible) ROI of this effort.
Having worked in Operations Leadership for over 13 years, if a consultant tried to explain this to me as a metric for adoption, I would have stopped them by the second variable explaining front line employees – the true brand ambassadors of any organization – would not understand it.
I recommend that you stick with just the NPS – it's not perfect, but it's easy to understand by everyone and easily compliments hardcore metrics. Don't try to reinvent the wheel with these efforts.
Cheers,
Mark
I'm very uneasy about both TSCV and SIM.
Jeremiah – you present TSCV as an “accurate” method for measuring customer value on the social web, but provide no justification for how the formula was constructed and the relationship between the variables. I'd love to see the formula applied with actual data from a real client, however I'm not sure that's possible. Reading through your comments to G2, you say “brands are rolling this out, let's see how it plays out”. This leads me to believe that what you are presenting as gospel is in fact a theory that lacks any historical data to substantiate. If this is still an unproven method, it's your duty, given your stature in social media, to point that out in your post.
Shiv – There is some lingering confusion over the formulas used to calculate SIM. Specifically, it's flawed in the reliance of industry net sentiment in determining your SIM score. The example posed on http://socialcommercetoday.com/social-media-met… takes a deep dive into SIM and uses real numbers to vet the formula. The results are a little disturbing.
I'm very uneasy about both TSCV and SIM.
Jeremiah – you present TSCV as an “accurate” method for measuring customer value on the social web, but provide no justification for how the formula was constructed and the relationship between the variables. I'd love to see the formula applied with actual data from a real client, however I'm not sure that's possible. Reading through your comments to G2, you say “brands are rolling this out, let's see how it plays out”. This leads me to believe that what you are presenting as gospel is in fact a theory that lacks any historical data to substantiate. If this is still an unproven method, it's your duty, given your stature in social media, to point that out in your post.
Shiv – There is some lingering confusion over the formulas used to calculate SIM. Specifically, it's flawed in the reliance of industry net sentiment in determining your SIM score. The example posed on http://socialcommercetoday.com/social-media-met… takes a deep dive into SIM and uses real numbers to vet the formula. The results are a little disturbing.
Companies don’t realize how important it is to monitor their social media through the use of a bi dashboard. Â The social web has a large impact on the purchases of the consumer.Â
Companies don’t realize how important it is to monitor their social media through the use of a bi dashboard. Â The social web has a large impact on the purchases of the consumer.Â